| Summary: | 碩士 === 國立臺灣海洋大學 === 航運管理學系 === 101 === The purpose of this research is to make comparisons on five different univariate forecasting methods, and find the best forecasting methods for the price of the Solar PV inverter, to forecast the monthly variations of the Solar PV inverter’s price. The data were collected from January 2007 to December 2011. Through the use of five different forecasting methods, namely Classical Decomposition Model, the Trigonometric Model, the Regression Model with Seasonal Dummy Variables, the Grey Forecast, and the Hybrid Grey model, we forecast the price of the solar PV inverter. The contribution of this research is to compare the forecasting results of the five univariate methods based on commonly used evaluation criteria, Mean Absolute Error, Mean Absolute Percent Error and Root Mean Squared Error, and to find out which can provide the most accurate predictions.
This research shows that the Hybrid Grey Model provided the most accurate predictions; Classical Decomposition Model is the next best. In conclusion, the Hybrid Grey Model provided the most accurate predictions for the short term demands for the price of the Solar PV inverter and Classical Decomposition Model is the runner up. Lastly, the outcome of this work can be helpful to find the best method to provide the most accurate predictions, as a reference for future Solar PV inverter purchase, sales, management, and research, as well as a reference to green energy industry.
Due to limited models used in our research, in the future we should consider to explore other forecasting models such as neural networks, AI methods and other forecasting methods to obtain a more accurate prediction for the price of the Solar PV inverter. Furthermore, research can be focused on prices for other products; such as: Cell, Poly, and Module, to verify models.
|
|---|
